Five false threshold tables in Cohen's conjectures on primes and cyclic numbers
Abstract
A positive integer n is cyclic if every group of order n is cyclic, equivalently (Szele) if gcd(n,φ(n))=1. Cohen's Conjectures about primes and cyclic numbers collects sixty-three conjectures: cyclic-number analogues of classical conjectures about primes, together with new conjectures about the primes themselves. Many of them have a two-part shape: an asymptotic existence clause "for every k there is a threshold N(k) such that …", followed by a printed table of specific values of N(k). The existence clauses are out of reach; the tables are finite claims. We show that five of those tables are false: Conjectures 17, 20, 22, 28 and 29 fail at thirty-nine of their printed constants, and we exhibit an explicit witness for each. Four of the five refutations follow from counts that the source prints itself, in one case three lines below the conjecture they contradict, and in another from a sequence in the OEIS that predates the conjecture by nine years; we recompute all of them from the definitions rather than quoting them. The refutation of Conjecture 22 is horizon-free: because its threshold is defined to be least, one count at one index suffices, with no claim about large arguments. We also show that three further printed thresholds can be lowered unconditionally. What is not refuted, in every case, is the existence clause; we separate the two halves explicitly, because the recent literature reports Conjectures 17 and 20 as proved, and the only statement that literature prints for either conjecture is the existence clause alone. All statements are verified in Lean 4 by the kernel, with no compiled evaluation anywhere.
Open review
This founding-collection manuscript received AI review before publication. Independent human review is open. Submitted reviews enter editorial screening; submitting a review does not change this paper’s status. Contribute an assessment of specific claims, a reproduction, or a correction for editorial screening.
Archived files
- Version 1 · current (opens in a new tab)
Source snapshot 2026-08-30 15:34 UTC
File fingerprint
8e4653be45ca40ca706f57a1926e4bd9ee5a747ab84d2f9d9213721dfb9c5177
Claim ledger
Stated results
ct-01candidate2026-08-28
Cohen's Conjecture 17 is false as printed: ten of its twelve tabulated thresholds fail, the cleanest at k = 2, where [289, 306] holds exactly one prime
ct-02candidate2026-08-28
Cohen's Conjecture 20 is false as printed: eleven of twelve tabulated thresholds fail, and the k = 2 cell contradicts the paper's own table three lines below it – under both readings of its printed 'prime'/'cyclic' typo
ct-03candidate2026-08-28
Cohen's Conjecture 22: fourteen of the seventeen printed B(k) are not the least thresholds his own definition names – refuted horizon-free, with no claim about large m
ct-04candidate2026-08-28
Cohen's Conjecture 28 is false as printed: N(10) = 11 fails at n = 11, where (11³, 12³) holds nine pairs of twin primes – as the paper's own printed table says
ct-05candidate2026-08-28
Cohen's Conjecture 29 is false as printed: N(8) = N(9) = N(10) = 12 all fail at n = 12, where (12³, 13³) holds seven pairs of cousin primes
ct-06candidate2026-08-28
Three further printed thresholds of Conjectures 28 and 29 are provably not least, by unconditional one-step shifts
ct-07routine2026-08-28
Each printed conjecture splits as (existence clause) AND (table of constants), and only the table is refuted – the separation that distinguishes these results from the literature's claim that Conjectures 17 and 20 are proved
ct-08routine2026-08-28
Compute-first-values gate: Cohen's five printed data tables recomputed from the definitions, kernel-checked
ct-09routine2026-08-28
Cyclic-number validation anchors, including Ibarra's kernel-checked Conjecture 66 counterexample re-derived from this family's own definitions
ct-10routine2026-08-28
Negative controls: too-large and too-small at the headline cell, the trial-division bound, the consecutiveness clause, and Cohen's inequality conventions
Provenance
- Generated by
- Machina Mathematica
- Released by
- Korea Superintelligence Labs
- Source context
- J. E. Cohen, *Conjectures about primes and cyclic numbers*, arXiv:2508.08335v1 = *J. Integer Seq.* 28 (2025), Article 25.4.7 (https://cs.uwaterloo.ca/journals/JIS/VOL28/Cohen/cohen41.pdf, published 7 Aug 2025).
- Snapshot
- 2026-09-07 03:53 UTC
- Ledger commit
801848d7